April 30 news, according to the "net letter China" public news today, in order to regulate AI services and applications, promote the healthy and orderly development of the industry, and safeguard the legitimate rights and interests of citizens, recently, the CentralInternet Information OfficeIssuance of circulars.Deploying a 3-month nationwide campaign called "Clearing the Air: Rectifying the Misuse of AI Technology"..
The relevant person in charge of the Central Internet Information Office said thatThis special operation is carried out in two phasesThe first stage is to strengthen the governance of AI technology at the source. In the first phase, we strengthened the governance of AI technology at the source, cleaned up and remedied illegal AI applications, strengthened the management of AI generation and synthesis technology and content labeling, and promoted websites and platforms to enhance their ability to detect and identify counterfeits. In the second phase, we will focus on the use of AI technology to produce and publish rumors, inaccurate information, pornographic and vulgar content, impersonate others, engage in network army activities and other prominent issues, focus on cleaning up relevant illegal and undesirable information, and dispose of and penalize non-compliant accounts, MCN agencies and websites and platforms.
1AI attached the specifics below:
Phase IFocusing on six outstanding issues:
- One is the offending AI product.Using generative artificial intelligence technology to provide content services to the public, but not fulfilling the large model filing or registration procedures. Providing functions that violate laws and ethics, such as "one-click undressing". Cloning or editing other people's voices, faces and other biometric information without authorized consent, violating other people's privacy.
- The second is to teach and sell tutorials and merchandise for offending AI products.Information on tutorials that teach the use of illegal AI products to fake face-changing videos, voice-changing audio, and other tutorials. Information on selling illegal "voice synthesizers", "face-changing tools", and other products. Information on marketing, hype, and promotion of the offending AI products.
- Thirdly, the training corpus is poorly managed.Use of information that violates the intellectual property rights, privacy rights, and other rights of others. Use of false, invalid, or inaccurate content crawled from the Internet. Using data from illegal sources. Failure to establish a training corpus management mechanism, and failure to regularly investigate and clean up violating corpus.
- Fourth, security management measures are weak.Failure to establish security measures such as content review and intent recognition that are appropriate to the scale of the business. Failure to establish an effective management mechanism for violating accounts. Security self-assessment has not been carried out on a regular basis. Social platforms are not clear about the number of AI auto-replies and other services accessed through API interfaces, and are not strict in gate-keeping.
- Fifth, the content labeling requirements have not been implemented.The service provider fails to add implicit and explicit content identification to deeply synthesized content, and fails to provide or prompt users with explicit content identification functions. The content dissemination platform did not carry out monitoring and screening of generated synthetic content, resulting in false information misleading the public.
- Sixth, security risks in key areas.Where AI products have been filed to provide Q&A services in key areas such as healthcare, finance, minors, etc., they have not set up targeted security audits and control measures for the industry, and problems such as "AI prescription", "induced investment", "AI illusion", etc. have appeared. AI prescribing", "inducing investment", "AI hallucination" and other issues, misleading students and patients, and disrupting the order of the financial market.
Phase IIThe focus is on seven outstanding issues:
- One is the use of AI to create and publish rumors.Creating rumors out of thin air, fabricating information about current politics, public policies, social and livelihood issues, international relations, emergencies, etc., or making unauthorized speculations or malicious interpretations of major guidelines and policies. To make up or fabricate reasons, progress and details of emergencies, disasters and accidents. Impersonating official press conferences or news reports to release rumor information. Utilizing content generated by AI cognitive bias for malicious guidance.
- The second is the use of AI to create and publish inaccurate information.Piecing together and editing unrelated graphics and videos to generate mixed and half-true and half-false information. Fuzzy modification of the time, place, characters and other elements of the event, rehashing old news. Produce and publish exaggerated, pseudo-scientific and other inaccurate content related to finance and economics, education, justice, healthcare and other specialized fields. Use AI fortune-telling and AI divination to mislead and deceive netizens and spread superstitious ideas.
- Third, the use of AI to produce and distribute pornographic and vulgar content.Use AI stripping, AI drawing and other functions to generate synthetic pornographic content or other people's indecent pictures and videos, soft pornography such as revealing clothes, head scratching, and other secondary marginal images, or play ugly wind and other undesirable oriented content. Produce and publish bloody and violent scenes, human body distortion, surreal monsters and other horror and bizarre images. Generate and synthesize novels, posts and notes with obvious sexual connotations, such as "pornography" and "dirty jokes".
- Fourth, the use of AI counterfeit others to commit infringement of the law.Through AI face-swapping, voice cloning and other in-depth forgery techniques, they impersonate experts, entrepreneurs, celebrities and other public figures to deceive netizens and even make profits through marketing. Using AI to spoof, discredit, distort and alienate public figures or historical figures. Using AI to impersonate friends and relatives and engage in illegal activities such as online fraud. Inappropriately using AI to "resurrect the dead" and misuse the information of the dead.
- Fifth, the use of AI to engage in cyber-mercenary activities.Using AI technology to "raise numbers", simulating real people to register and operate social accounts in bulk. Using AI content farms or AI manuscript washing to generate and publish low-quality and homogenized texts in bulk to gain traffic. Use AI group control software and social robots to batch like and comment on posts, brush volume and control comments, and create hot topics on the list.
- Sixth, AI product service and application violations.Production and dissemination of counterfeit or shell AI websites and applications, AI applications providing illegal functions, such as creation tools providing functions like "expanding hotspots on hot search lists into articles", and AI social networking and chatting software providing vulgar and soft-pornographic dialog services. Providing illegal AI applications, generating synthesis services or courses for sale or promotion to attract traffic.
- (vii) Violation of the rights and interests of minors.AI apps inducing minors to become addicted, the presence of content affecting minors' physical and mental health in the minors' mode, and so on.
The person in charge of the Central Internet Information Office emphasized that local Internet information departments should fully understand the significance of the special action to prevent the risk of AI technology abuse and safeguard the legitimate rights and interests of Internet users. Effectively fulfill the responsibility of local management, supervise websites and platforms to improve the review mechanism of AI-generated and synthesized content against the requirements of the special action, improve the technical detection capability, and do a good job of rectification and implementation. Strengthen the publicity and promotion of AI-related policies and popularization education on AI literacy, guide all parties to correctly understand and apply AI technology, and continuously build consensus on governance.